An interactive analysis and display system (ADIEEG) has been implemented to perform parallel spectral and transient analysis and display in real time or faster on up to 16 EEG channels. Among its characteristics are automatic rejection of gross EEG artifact and detection of sharp transient activity characteristic of the epilepsies. Thresholded linear coherence spectra are used to quantify asymmetries and foci associated with structure-altering processes, the epilepsies, and other pathological conditions. The proposed renewal of this project would allow further development and evaluation of the metrics, algorithms and program modules essential to quantifying and automating clinical EEG analysis, culminating in the implementation and clinical evaluation of an inexpensive microcomputer-based analysis system. Essential to this effort is the continued assembly of a standardized clinical EEG data base. Objective evaluation of improvements in automated EEG classification performance resulting from heuristic feature extraction will be made. EEG metrics will be correlated with visual interpretations of the polygraph, the neurologic examination and other standard tests. A series of two-class decision rules will be derived and incorporated into the on-line system, allowing analysis to both track significant electrophysiologic characteristics and terminate itself when classification can reasonably be made. The data base, primary analysis, and feature extraction and classification program logic will be made available to other research groups working on these and related problems.